Table 3. Subgroup analyses: agreement between the health questionnaire and dietary history interview (n = 4254).

	n	ra	DHIb (grams/day)	diff.c (%)	Pd
Age (years)
20–29	26	0.36	722	8.0	0.45
30–59	3379	0.87	641	0.4	0.41
≥60	849	0.80	570	−2.1	0.06
Sex
Male	2194	0.85	641	0.7	0.31
Female	2060	0.88	612	−0.8	0.14
Marital status
Married	3360	0.86	635	0.006	0.99
Living alonee	894	0.87	598	−0.1	0.92
Education (years)
1–9	3140	0.85	643	−0.4	0.44
≥10	1114	0.89	583	1.1	0.16
Community density
Rural	1182	0.84	645	−2.3	0.006
Urban	3072	0.87	620	0.9	0.07
Alcohol consumption (grams/day)
0	1773	0.86	617	−1.4	0.02
<12	1852	0.88	645	0.8	0.17
≥12	629	0.82	605	1.5	0.26
Body mass index (kg/m2)
<25	1774	0.85	626	−0.5	0.49
25–29.9	1841	0.88	629	0.4	0.54
≥30	639	0.84	626	0.1	0.90
Physical activityf
None	1079	0.82	650	0.3	0.75
Light activity ≥4 hours/week	2734	0.86	625	−0.2	0.71
Heavy activity ≥3 hours/week	410	0.92	589	0.3	0.80
Heavy activity almost daily	31	0.98	536	0.7	0.77
Smoking
Never	2332	0.89	584	−0.5	0.27
Past	860	0.85	619	0.4	0.70
Current	1062	0.81	730	0.6	0.50
Hypertension
No	3170	0.86	639	0.2	0.74
Yes	1084	0.86	593	−0.6	0.48
Serum cholesterol
<7.008 mmol/lg	2123	0.87	610	0.1	0.82
≥7.008 mmol/l	2131	0.86	645	−0.2	0.78

^aIntraclass correlation coefficient, coffee consumption measured as a continuous variable grams/day.

^bDietary history interview, average daily coffee consumption in grams.

^cDifferences between the health questionnaire and dietary history interview (%), where negative values indicate that reported coffee consumption was higher on the health questionnaire.

^dStatistical significance of the difference between the 2 methods.

^eSingle, divorced, or widow/er.

^fLeisure-time physical activity.

^g7.008 mmol/l was the median serum cholesterol level of the subjects.